"With an annual budget of about $8.3 billion (fiscal year 2020), the NSF funds approximately 25% of all federally supported basic research conducted by the United States' colleges and universities.[4] In some fields, such as mathematics, computer science, economics, and the social sciences, the NSF is the major source of federal backing." - Wikipedia

Let's say that the NSF, during a typical fiscal year, received requests for research funding which totaled far more dollars than the NSF actually had available to grant. And that the NSF had a long-established system of paperwork, procedures, and preferences - to decide which researchers would receive how much from the NSF's limited funds.

Now, if the NSF's preferences were inclined against supporting closed-access publication of research results, and if some universities were not so flush with research funding from other sources that they could happily maintain a "principled stance" and ignore the NSF's crass, self-serving bias...